Lecture 12 exceptions

Introduction to
computational thinking
Module 12 : Exceptions
Asst Prof Michael Lees
Office: N4‐02c‐76
email: mhlees[at]ntu.edu.sg
1

Contents
1. Why & What are exceptions
2. Try/Except group
3. Exception Philosophy
4. else and finally suites
Chapter 14
2

Dealing with problems
• Most modern languages provide ways to deal
with ‘exceptional’ situations
• Try to capture certain situations/failures and
deal with them gracefully
• All about being a good programmer!
Module 12 : Exceptions 3

What counts as an exception
• Errors: Indexing past the end of a list, trying to
open a nonexistent file, fetching a nonexistent
key from a dictionary, etc.
• Events: Search algorithm doesn’t find a value (not
really an error), mail message arrives, queue
event occurs.
• Ending conditions: File should be closed at the
end of processing, list should be sorted after
being filled.
• Weird stuff: For rare events, keep from clogging
your code with lots of if statements.

General idea
4 general steps:
1. Keep watch on a particular section of code
2. If we get an exception, raise/throw that
exception (let it be known)
3. Look for a catcher that can handle that kind
of exception
4. If found, handle it, otherwise let Python
handle it (which usually halts the program)

Bad input
• In general, we have assumed that the input
we receive is correct (from a file, from the
user).
• This is almost never true. There is always the
chance that the input could be wrong.
• Our programs should be able to handle this.
• "Writing Secure Code,” by Howard and LeBlanc
– “All input is evil until proven otherwise.”

General form v1
try:
Code to run
except aParticularError:
Stuff to do on error

Try suite
• The try suite contains code that we want to
monitor for errors during its execution.
• If an error occurs anywhere in that try suite,
Python looks for a handler that can deal with
the error.
• If no special handler exists, Python handles it,
meaning the program halts and with an error
message as we have seen so many times 
try:
Code to run
8

Except suite
• An except suite (perhaps multiple
except suites) is associated with a try
suite.
• Each exception names a type of exception it is
monitoring for (can handle).
• If the error that occurs in the try suite
matches the type of exception, then that
except suite is activated.
9

try/except group
• If no exception in the try suite, skip all the
try/except to the next line of code.
• If an error occurs in a try suite, look for the
right exception.
• If found, run that except suite, and then
skip past the try/except group to the next
line of code.
• If no exception handling found, give the error
to Python.

An example
try:
print('Entering try suite') # trace
dividend = float(input('dividend:'))
divisor = float(input('divisor:'))
result = dividend/divisor
print('{:2.2f} divided by {:2.2f} = ’
'{:2.2f}'.format(dividend,divisor,result))
except ZeroDivisionError:
print('Divide by 0 error')
except ValueError:
print("Couldn't convert to a float")
print('Continuing with the rest of the program')
Try Suite
Except
Suite 1
Except
Suite 1
Try/Except Group
12

What exceptions are there?
• In the present Python, there is a set of
exceptions that are pre‐labeled.
• To find the exception for a case you are
interested in, easy enough to try it in the
interpreter and see what comes up.
3/0 =>
ZeroDivisionError: integer division or modulo by
zero

Details: http://docs.python.org/library/exceptions.html
14

EXCEPTION PHILOSOPHY

How you deal with problems
Two ways to deal with exceptions:
• LBYL: Look Before you Leap
• EAFP: Easier to Ask Forgiveness
than Permission (famous quote
by Grace Hopper)

Look before you leap
• Super cautious!
• Check all aspects before executing
– If string required : check that
– if values should be positive : check that
• What happens to length of code?
• And readability of code – code is hidden in
checking.

Easier to ask forgiveness than
permission
• Run anything you like!
• Be ready to clean up in case of error.
• The try suite code reflects what you want to
do, and the except code what you want to
do on error. Cleaner separation!

A Choice
• Python programmers support the EAFP approach:
– Run the code (in try) and used except suites to deal
with errors (don’t check first)
if not isinstance(s, str) or not s.isdigit:
return None
elif len(s) > 10: # too many digits to convert
return None
else:
return int(str)
try:
return int(str)
except (TypeError, ValueError, OverflowError):
return None
LBYL
EAFP
19

OTHER SUITES

else suite
• The else suite is used to execute specific code
when no exception occurs
try:
Code to run
else
Stuff to do when no error
21

finally suite
• The finally suite is used to execute code at the
end of try/except group (with or without
error)
try:
Code to run
finally
Stuff to do always at end
22

Challenge 12.1 Calculation errors
Modify the calculator example from module 4 to capture some common
exceptions.
ERROR!
23

Thought process
• What errors can the code generate?
• Capture the common ones and give clear
instruction
• Use the else condition to print out answer
only when no error

Take home lessons
• Using exception handling to help users is
important!
• Different types of exceptions: events, errors,
etc.
• LYBL vs. EAFP
• try‐except‐else‐finally suites (in many
languages)

Further reading
• http://docs.python.org/tutorial/errors.html
• http://diveintopython.org/file_handling/index
.html
• http://oranlooney.com/lbyl‐vs‐eafp/

Lecture 12 exceptions

Recomendados

Recomendados

Más contenido relacionado

Más de alvin567

Más de alvin567 (13)

Último

Último (20)

Lecture 12 exceptions